Back

Modelling the future distribution of rare bryophytes in Scotland: is inclusion of habitat loss important?

Ferretto, A.; Smith, P.; Genney, D.; Matthews, R.; Brooker, R.

2021-08-31 ecology
10.1101/2021.08.30.458156 bioRxiv
Show abstract

Species distribution models (SDMs) have been widely used to predict species ranges and their future distribution under climate change scenarios. In this study we applied Maxent, one of the most used SDMs, to project the distribution of some rare bryophyte species in Scotland in the 2050s. Most of these species are strongly linked to the blanket bog habitat, which is threatened by climate change in the near future. To assess the extent to which changes in habitat distribution leads to a different modelled distribution of the selected bryophytes, blanket bog distribution was included in the model as one of the explanatory variables for some species, and Maxent was run for three 2050s scenarios: once with the current blanket bog distribution and two other runs using the blanket bog distribution derived from two bioclimatic models (Lindsay modified and Blanket Bog Tree model) under the same climate change scenario. For seven out of nine of our studied bryophyte species, the modelled distribution in Scotland was predicted to decline, with some species retreating towards the north-west and other species almost disappearing. When the change in blanket bog distribution was also accounted for, further areas in the north/centre east of Scotland and in the south were predicted to be unfavourable for many of the species considered. Our findings suggest that when modelling species distributions, habitat distribution also needs to be considered, especially when there is a strong relationship between the species and a particular habitat.

Matching journals

The top 7 journals account for 50% of the predicted probability mass.

50% of probability mass above

"Similar papers" are the closest papers from that journal in the model's embedding space. They show what the match is built on, but the ranking comes mostly from a classifier over the whole training set, not from these examples alone.